Table 2.

Key differences between midterm and endline surveys

	Midterm	Endline
Objective	Measure mortality change between preimplementation to postimplementation phase	Measure mortality change between midterm and endline
Survey design	Multistage cluster survey with a semirandomized, stepped wedge component	Multistage cluster survey
Sample size	6,067 households, 96 clusters	5,125 households, 90 clusters
Stratification	PHU	Small, medium, large villages
Sampling	Sampled urban areas (excluded in the analysis)	No urban areas sampled
Data collection	Respondents referred to treatments using a picture card	Respondents referred to treatments using samples of actual treatments (bottles, blister packs)
	Pneumonia case definition required that respondents reported fast or difficult breathing	Pneumonia case definition modified so that if a cough/fast or difficult breathing/cold were reported, respondents were prompted about the other symptoms